I made some mistakes in my last post. First off, I’d like to apologize to Ashe Dryden for referring to her as Ms. Dryden while I referred to James Coglan by his full name. I had no intention of explicitly implying Ashe’s gender; my usual writing style is to refer to people by their full name initially and to use an approrpiate title thereafter. But upon review I realized I referred to Ashe as “Ms. Dryden” in subsequent references, while I referred to James Coglan as “James Coglan” both times I mentioned him.
I’m sorry about that. It was an oversight, and if I had been paying better attention, the second reference to Mr. Coglan would have been “Mr. Coglan” for parity. But as I failed to achieve parity, and Ashe expressed that she preferred “Ashe Dryden” to “Ms. Dryden” in general, I decided to
s/Ms./Ashe/g the post. It is now corrected.
I missed some things
My post was mostly in response to the Github Is Not Your CV sentiment, as well as the Hacker News comments related to Ashe’s post, rather than the diversity/privilege statistics offered by Ashe’s original article. I had no intention of discounting the privilege issues involved in having the time and resources to contribute to OSS or produce original experiments/libraries/applications on Github.
I didn’t disagree with any of that. I still think it’s a good idea to make projects public and make that part of your portfolio. I never suggested that hiring managers should use Github as the only criterion… but I do think it demonstrates a commitment to professional development, to honing your craft.
I’m curious if the cited statistics on OSS contributions refer to “have commits to major OSS projects” or “have code samples online.”
My reference was to the latter. Not that I wouldn’t love to hire people with commits to Rails or core.async; but for the most part, I want to see what you build in your spare time. Your production code is probably proprietary and covered by an NDA. So I want to see what I can see.
If you don’t have a Github profile, that’s okay. We can still talk. Either way, I’m probably going to want to pair with you or give you a take-home assignment before the next interview round. Those are great ideas that both Ashe and James proposed as alteratives to Github reviews.
I see them more as complements than alternatives, but ideally seeing both would be wonderful.
I still love Github as part of the screening process. Especially if your Github is messy. I can see the crappy Rails CMS you wrote two years ago as practice, and the Clojure OAuth library you comitted to last week, and see the growth you’ve made in that time. That’s the value of Github to me.
I want to see professional development, not unpaid labor. I want to know you don’t consider hacking labor, but play. Not because I want you to work 80 hours a week; I don’t, because that’s the road to burnout. But I want to know you love this shit and, in a socialist utopia, would do it for free.
Yeah, I know hacking on OSS is a privilege we don’t all enjoy. Neither is going to college or learning to code. These are all things we need to improve access to as a community, but the nice thing about Github, pairing, and code challenges, is that they can balance out hte privilege issues of college and other factors. For me, as a hiring hacker, I care more about work than degrees. Github is one way to show that, but pairing etc. are equally valid.
OSS hacking is a privilege: I’ve done a fuckton less since I had a kid; mostly work-related stuff. But it still matters, and it’s accessible to anyone with a computer and an internet connection. That in and of itself is discriminatory, but I can’t see how you’re going to be a hacker to begin with sans those things.
I think my main objection is the Harrison Bergeron argument: we’re using statistics about privilege to say “let’s not use this (IMHO) valid metric because the privileged have access to this resource” instead of “how can we change society to improve access to this resource.”
Yeah, the second proposition is harder. And you can discount the utility of the metric like James Coglan, or reject the premise like Ashe Dryden, but in either case there is a realpolitik question:
- Hiring managers see value in this metric.
- Some of them probably have valid reasons for this.
- Access is influenced by social privilege.
- We can either (a) throw the baby out with the bathwater, or (b) try to address the underlying issue.
The typical response is 4(a), for whatever reason. I’d like to see 4(b), but I don’t know how to make it happen. I’d rather see theories on how this can be achieved than see strategies and arguments for 4(a).
Anyhow, like I said… the statistics are damning, but I’d rather see the statistics change. How do we get more women into open source? I think OSS contributions have value as one factor among many, but I also think it’s a value for the individual developer, not just as a hiring criterion. The low contribution ratio of women and minorities to OSS is the problem, not the fact that it’s a metric used for candidate evaluation. But that’s a whole different issue, and I don’t have an answer for fixing that… yet.
tl;dr apologies to Ashe; inequality and privilege are real; fix society, don’t discount the value of OSS/Github contributions